Aspects of Copper Price Volatility

ICSG Seminar, Lisbon, October 2007 Alan Williamson

alan.williamson@trafigura.com +44 207 009 1702

Trafigura
Trafigura is a USD50Bn a year commodity trader Second largest independent non-ferrous metals trader Third largest independent oil trader 50 offices in over 35 countries Galena Asset Management Subsidiary of Trafigura Leveraging off group information Proprietary research into supply and demand 3 Main Funds, cUSD600m under management

Trafigura 2007

Structure of presentation
Recent price history Who are the funds involved in the commodity markets? How do we track what the funds are doing? What is the impact of fund activity on metals prices and volatility? Other factors affecting prices and volatility
The US dollar Arbitrage between the LME and other exchanges The role of non-visible stocks

Summary and conclusions

Trafigura 2007

Long run trends in copper prices

In nominal terms current copper prices are unprecedented but in real terms we have been here before The surge in prices 2002-07 has been unusual (outside war)
c/lb 500 400 300 200 100 0 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Nominal and real copper prices

USD/tonne 10000 8000 6000 4000 2000 0

Nominal price Source: CRU, Trafigura

Trafigura 2007

Real copper price (2007 money )

Why have prices surged so dramatically?

%yoy 7 6 5 4 3 2 1 0 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 average = 3.8%

Global GDP growth has been exceptionally strong

forecast

The supply side has struggled

Mt 18 17 16 15 14 Feb-05 Apr-05 Nov -05 Apr-06 Oct-06 Actual

ICSG forecasts for 2006 copper mine output

weeks 7 6 5 4 3 2 1 0 Jan-90

Stocks have been depleted

USD/t

10000 6000 4000 2000 0 Jan-93 Jan-96 Jan-99 Jan-02 Jan-05

5

160 140

and the funds have poured in

USD Billions

8000

120 100 80 60 40 20 0 1999 2000 2001 2002 GSCI-DJ-AIG 2003 2004 2005 2006 2007e Other indices

Stocks:Con Ratio
Trafigura 2007

LME Cash

Types of funds involved in the metals markets

Hedge funds
Over 10,000 active funds (more than the 8,600 US mutual funds) Current assets around USD1,500Bn (US mutual funds cUSD9,200Bn) Have a variety of investment styles, across asset classes Can be long or short

Commodity Trading Advisors

Current assets cUSD180Bn Are largely driven by technical factors (charts, momentum) Can be long or short

Pension funds
Traditionally were purely index investors (GSCI, for example) Not particularly price sensitive Almost always long

Other (arbitrageurs, specialists, Chinese general public)

Trafigura 2007

Growth in hedge funds has been enormous

Hedge fund business has increased 10x in recent years Over 10,000 hedge funds with cUSD1,500Bn under management Main sources of funding are high net worth individuals and Fund of Funds
Number 12000 10000 8000 6000 4000 2000 0 1992 1994 1996 1998 2000 2002 2004 2006 Assets under management (RHS) Hedge funds (LHS) USD Billions 1600 1400 1200 1000 800 600 400 200 0

Trafigura 2007

Hedge fund strategies

Arbitrage seek to aim from asset mis-pricing
Convertible arbitrage between bonds and equities Fixed income arbitrage between bonds of different maturities, grades etc Risk arbitrage where different assets are pricing in different risk probabilities Derivative arbitrage between a derivative and the underlying asset

Emerging markets - buy equity/debt in emerging markets Event driven

Distressed securities - buy debt/equities at discount Merger arbitrage between acquiring company and target company

Global macro aim to profit from changes in global economy Fund of funds - mixes and matches hedge funds Income - primary focus is yield Long/short equity - low volatility by offsetting positions across markets Market timing - attempts to pre-empt changes in market/economic conditions Multi-strategy - mixes and matches different hedge fund strategies
Trafigura 2007

Characteristics of macro hedge funds

Aim to benefit from changing global economy Can take positions across asset classes, but often specialise Use derivatives to gain leverage and maximise gains Are very well researched Look for various investment opportunities Cyclical play (near the top or bottom of a cycle)
Changing fundamentals / end-use applications Look for value in commodity / equity trades

Can run positions / trades for very long, or very short time periods

Trafigura 2007

Macro funds can shorten bear markets

In traditional cycles (the 1980s) inventories needed to fall to critical levels before prices would take off This changed in the 1990s as funds moved to anticipate the fundamentals This ensured the bear market ended sooner than it otherwise would have done Risk/reward means this is more noticeable in bear markets than bull markets
weeks 40 35 30 25 20 15 10 5 0 -5 -10 1980
Trafigura 2007

"Excess" metal inventories and prices

price rise on low inventories price rise despite high inventories

Index 200 180 160 140 120 100 80 60 40

1983 1986 1989 1992 1995 Excess inventories - all metals (LHS)
10

1998

2001 2004 2007 Base metal prices (RHS)

Commodity Trading Advisors (aka Managed Futures)

Like hedge funds the growth in CTA business has been tremendous Long term average returns of CTAs have been 13.6%pa Little correlation with equities or bonds
USD Billions 200 180 160 140 120 100 80 60 40 20 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 Assets under management
Trafigura 2007

11

Characteristics of CTAs
Can be technically driven or discretionary
90% technically driven (charts, momentum) 10% discretionary

Allocate across asset classes

Foreign exchange 40% Financials 30% Softs 14% Metals 9% Energy 7%

Allocation across LME metals

Al 41%, Cu 27%, Zn 15%, Ni 10%, Pb 4%, Sn 3%

Account for 15-20% LME turnover

Trafigura 2007

12

CTAs in the copper market??

2005 to May 2006 CTAs long as prices supported by 40 day moving average Momentum wanes in summer 2006 and CTAs switch to short positions Early 2007 momentum turns higher and CTAs switch back to long positions Currently still long
9000 8000 USD/tonne 7000 6000 5000 4000 3000 2000 Jan-05 Jul-05 LME copper Jan-06 Jul-06 40 day moving average
13

Jan-07 Jul-07 200 day moving average

Trafigura 2007

Pension fund interest in commodities

Few pension funds have explicit mandates to invest in commodities, but this is growing
Dutch pension fund PGGM in 2000, now 4% of USD76Bn funds Stichting Pensioenfunds in 2001, now 2.5% of USD190Bn funds Ontario Teachers, Missouri State, Harvard University, Novartis have followed

Investment decisions are based principally on asset class allocation considerations not supply/demand
Long run rates of return equal to or greater than other assets Low correlation with equities and / or bonds Returns/investment decision can be influenced by what happens in energy

Are generally passive investors: historically have invested in index products or fund of funds
Increasingly looking to enhance returns

Trafigura 2007

14

How much is invested in the commodity sector?

Fund interest has boosted commodity prices and increased volatility Best estimates suggest USD145Bn in index products, USD40-50Bn in base metals No immediate sign of investor appetite for metal exposure waning as a result of US sub-prime worries Continues trend to more active management
160 140

Investment in commodity indices

cUSD50Bn spec. money in base metals in 2007

Commodity index funds, 27%

USD Billions

120 100 80 60 40 20 0 1999 2000 2001 2002 GSCI-DJ-AIG 2003 2004 2005 2006 2007e Other indices

CTAs and Prop Desks, 36%

Hedge funds, 37%

Trafigura 2007

15

How do we track fund activity?

Very little hard data available LME open interest
Rising open interest and rising prices = fund long position building Falling open interest and falling prices = fund long liquidation Rising open interest and falling prices = fund short position building Falling open interest and rising prices = fund short covering But other factors (producers/consumers) also impact the open interest data

CFTC Commitments of Traders reports

Comex copper, aluminium, gold and silver Nymex platinum and palladium

Anecdotal evidence the funds are buying . the funds are selling Replicate the funds run simple technical models
16

Trafigura 2007

LME copper open interest and prices

'000 lots 260 250 240 230 220 210 Jan-06

LME copper open interest and prices

USD/tonne 9000 8000 7000 6000 5000 4000

May-06

Sep-06

Jan-07

May-07

Sep-07

LME copper open interest

LME cash copper

Trafigura 2007

17

CFTC Commitments of Traders data

Most analysts follow the CFTC CoTs data The non-commercial + non-reportable positions are considered a proxy for speculative involvement in the markets But there are huge problems with the data that make it difficult to read, particularly for copper
'000t 700 600 500 400 300 200 100 0 -100 -200 -300 Jan-02 c/lb 410 360 310 260 210 160 110 60 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Copper price (3mths) - RHS

Speculative position on Comex - LHS

Trafigura 2007

18

Influence of funds on prices

Macro hedge fund Can be long or short (and sometimes long and short at the same time)
Positions often counter-cyclical (picking turning points) Effect is to dampen/shorten the cycle (particularly bear markets) No long term impact on prices

CTAs Can be long or short

Positions tend to follow market trends Effect is to exacerbate/lengthen the cycle No long term impact on prices

Pension Funds and other index investors Almost always structurally long
Effect is to raise long term prices Also has a significant impact on short term spreads as positions are rolled

Trafigura 2007

19

Index fund investment

S&P GSCI has cUSD80Bn invested in commodities Currently copper has a 3.98% weighting
= USD3.18Bn = c400,000t copper

DJ-AIG index has cUS40Bn invested in commodities Currently copper has a 6.19% weighting
= USD2.48Bn = 310,000t copper

Other indices have cUSD25Bn invested in commodities CRB has a 5.88% weighting in copper
S&P Index has a 3.39% weighting in copper Rogers International Commodity Index has a 4.00% weighting in copper Deutsche Bank LCI has a 0% weighting in copper = USD660Mn = 85,000t copper

Total index fund investment = USD6.32Bn = 795,000t copper

Trafigura 2007

20

Influence of index fund investment in copper

800,000t of long position building has undoubtedly boosted copper prices Quantifying the impact is difficult given other influences the level of stocks, strength of demand, other speculative interest in the market, the US dollar etc But clearly while index funds remain long, copper prices will be higher, for longer, than in previous cycles
USD/tonne 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 0
Trafigura 2007

LME copper stocks and 3 month prices 1994-2007

2006-7 Stocks.consumption relationship mov ing further out - higher price for a giv en stock lev el

2004-5 1994-2003

100

200

300

400

21

500

600

700

800

900

1000

LME Stocks '000t

 ditto aluminium
This structural shift higher in stocks/price relationships is common to all metals
USD/tonne 3500 3000 2006-7 2500 2000 1500 1000 0 500 1000 1500 2000 2500 3000 LME Stocks '000t 2004-5 1994 - 2003

LME aluminium stocks and 3 month prices 1994-2007

Trafigura 2007

22

A weak dollar is good for copper! Discuss

Conventional wisdom has it that a weak dollar is good for commodities This should be through marginal changes in supply and demand
A falling dollar lowers non-US prices, stimulating marginal demand, cutting production Lower supply and higher demand = rising US dollar denominated prices The impact should be higher with prices close to costs of production

The funds knowing of this relationship will buy in advance This relationship is particularly important for gold
Correlation coefficient weekly returns 2000-2007
0.00 -0.05 -0.10 -0.15 -0.20 -0.25 -0.30 -0.35
um ad er m ld nc ck e lv e Oi inu Ti Go in i
23

Si

Le

Co

Ni

at

um

Pl

Al

Pa

Trafigura 2007

ll a

d iu

Zi

pp

The role of the dollar in determining copper prices

The view that the dollar drives copper prices is unfounded empirically In 2002-3 copper and the dollar were positively correlated Since 2004 there has generally been a negative correlation, but is this causal or coincidental? If the copper market moves into oversupply expect copper to fall irrespective of what the dollar does
0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07

3m rolling correlation coefficient - copper vs the dollar

Trafigura 2007

24

The influence of multiple exchanges on price volatility

For many years the only copper arbitrage was between LME and Comex
Arbitrage difference was narrow as both contracts were liquid and nearby warehousing facilities ensures that metal could easily be moved to settle contracts Activity has declined as increased volatility has seen risk/reward deteriorate as has the decline in US copper consumption and production

This has now been surpassed by the LME-Shanghai Futures Exchange arbitrage
Wide arbitrage reflecting lack of close warehousing facilities (higher freight costs) exacerbated by Chinese import/export restrictions and taxes Making prices volatile due to a lack of clarity on why metal is flowing to China

Trafigura 2007

25

LME-SFE arbitrage

Shanghai prices are a function of LME prices, local supply and demand for physica metals, and local speculative activity This can be further affected by raw material prices and the forward structure on both exchanges Huge differences in LME and SHFE prices can develop that require physical shipment to correct
USD/tonne 11000 10000 9000 8000 7000 6000 5000 4000 Jan-06 SHFE cash
Trafigura 2007

LME and SHFE cash prices

Jul-06 Jan-07 Jul-07 LME cash SHFE cash post taxes

26

USD/tonne 1400 1200 1000 800 600 400 200 0 -200 -400 -600 -800 Jan-06

LME-SHFE arbitrage

Jul-06

Jan-07

Jul-07

Chinese imports and SFE-LME arbitrage

Early 2006 arbitrage was against imports weak demand and stock overhang Stocks were run down through H1 2006, causing the local market to tighten By late 2007 Shanghai was trading at a significant premium to LME Traders began shipping copper to China in record quantities Eventually pushing the market into oversupply, arbitrage disappeared Excess stocks now being worked off / exported / accumulated by SRB
$/tonne 400 200 0 -200 -400 -600 -800 -1000 2002
Trafigura 2007

Chinese market beginning to adjust

Shanghai at premium to LME

'000 200 180 160 140 120 100 80 60 40 20 0 2007

Shanghai at discount to LME

2003

2004

2005

2006

27 Monthly net refined copper imports

SFE-LME arbitrage

2006-07 movements in Chinese unreported stocks

2006 Chinese imports fell sharply despite ongoing increase in real demand

Chinese supplies of 3.58Mt and consumption of 3.95Mt implies stock drawdown of 370kt Market perceived lower imports to be bearish, erroneously believing demand to be weak

2007 Chinese imports rose sharply as favorable arbitrage encouraged shipment

2007 supplies of 4.75Mt and consumption of 4.50Mt implies stock build of 250kt Market perceived rising imports as bullish, erroneously believing demand to be stronger
500 400 '000tonnes Consumption 300 200 100 0 2000
Trafigura 2007

Chinese copper supplies and consumption

Stock building Destocking

2001

2002 2003 Refined output

28

2004

2005 2006 Net imports

2007

(Remember China is the key driver to copper prices)

For much of the last 3 years the markets have been sanguine about US growth, correctly being more interested in China This changed in August, the markets fixating on US debt problems This created some good buying opportunities, particularly in equities If US debt worries ease, expect attention to go back to China
US GDP growth and copper prices 2005-Q3 07
8000 7000 6000 5000 4000 3000 2000 1000 0 0 1 2 3 4 5 6 US quarterly GDP grow th % R 2 = 0.1061

Chinese GDP growth and copper prices 2005-Q3 07

9000 8000 7000 6000 5000 4000 3000 2000 1000 0 9 9.5 10

R 2 = 0.9203

10.5 11 11.5 Chinese quarterly GDP grow th %

Trafigura 2007

29

2007 Chinese stocks up, Chilean stocks down

H2 2006 collapse in US demand led to an unexpected rise in Chilean copper stocks Much of this was then exported in H1 2007, most to China Markets took this as bullish (lower surplus in H2 2006 and higher deficit in H1 2007 But should be market neutral as it is merely relocation of existing inventory
'000 tonnes 450 400 350 300 250 200 150 100 50 0 2003 2004 2005 Codelco strategic stockpile 2006 Other Chile 2007

Inferred Chilean stocks of copper back to normal

Stock change calculated as production less exports less consumption

30

Trafigura 2007

The role of stocks in determining copper prices

Basic economic theory has it that stocks influence prices But in reality information flows lag and it is difficult to know the true stock situation

390 340 290 240 190 140 90 40

Copper inventories and prices

current

USD/tonne
-300 -250 -200 -150

LME copper stocks and the cash-3m backwardation 1994-2007

USc/lb

R = 0.7218

-100 -50 0 50 0 100 200 300 400 500 600 700 800 900 1000 LME Stocks '000t

current

1 2 3 4 5 6 7 Total reported industry stocks (weeks consumption)

Trafigura 2007

31

But movements in unreported stocks can dwarf movements in reported stocks

Movements in copper stocks, 000t 2006 and 2007
Exchange stocks Other reported stocks Total reported stocks Chinese stocks Chilean port stocks Other unreported stocks Total other stocks Total stocks Market balance Prices H1 06 5 -16 -11 -210 -69 -100 -379 -390 Deficit Up H2 06 86 34 120 -170 168 100 98 218 Surplus Dow n H1 07 -21 -21 -42 250 -134 -50 66 24 Balanced Up H2 07e 20 0 20 0 40 40 80 100 Surplus Dow n

Trafigura 2007

32

Summary and conclusions

Copper prices have surged higher as the fundamentals have been incredibly tight The funds have had and will continue to have a huge impact on commodities But we need to differentiate between the funds macro funds, CTAs and long only index investors Macro and CTA influence is transient, long only investors should keep prices higher for longer The impact of the dollar in driving copper prices is overstated

The China factor production, consumption, stock movements, speculative activity and arbitrage is exacerbating copper price uncertainty and volatility

Trafigura 2007

33

Trafigura 2007

34